|Home | About | Journals | Submit | Contact Us | Français|
Vibrio cholerae ’s capacity to cause outbreaks of cholera is linked to its survival and adaptability to changes in aquatic environments. One of the environmental conditions that can vary in V. cholerae’s natural aquatic habitats is calcium (Ca+2). In this study, we investigated the response of V. cholerae to changes in extracellular Ca2+ levels. Whole-genome expression profiling revealed that Ca2+ decreased the expression of genes required for biofilm matrix production. Luria–Bertani (LB) medium supplemented with Ca2+ (LBCa2+) caused V. cholerae to form biofilms with decreased thickness and increased roughness, as compared with biofilms formed in LB. Furthermore, addition of Ca2+ led to dissolution in biofilms. Transcription of two genes encoding a two-component regulatory system pair, now termed calcium-regulated sensor (carS) and regulator (carR), was decreased in cells grown in LBCa2+. Analysis of null and overexpression alleles of carS and carR revealed that expression of vps (Vibrio polysaccharide) genes and biofilm formation are negatively regulated by the CarRS two-component regulatory system. Through epistasis analysis we determined that CarR acts in parallel with HapR, the negative regulator of vps gene expression.
Vibrio cholerae is a facultative human pathogen and the causative agent of the diarrhoeal disease cholera. The life cycle of this bacterium involves rapid growth within the human intestine and prolonged survival in aquatic habitats (Kaper et al., 1995; Faruque et al., 1998). Vibrio cholerae is a natural inhabitant of coastal and estuarine environments where it is found either as individual cells in the water column or as biofilm-associated cells attached to surfaces (Huq et al., 1983; Huq et al., 1986; Colwell and Huq, 1994; Huq et al., 1995; Islam et al., 2007). Indeed, V. cholerae’s ability to form biofilms is a key factor for its survival in its natural habitats and transmission to human host.
Production of mature biofilms in V. cholerae requires extracellular matrix components. The major component of the V. cholerae biofilm matrix is VPS (Vibrio polysaccharide) exopolysaccharide, and VPS production is essential for development of 3D biofilm structures (Watnick and Kolter, 1999; Yildiz and Schoolnik, 1999). The extracellular matrix of V. cholerae biofilms also contains proteins (RbmA, RbmC and Bap1) that stabilize the biofilm matrix, based on mutant phenotypes (Fong et al., 2006; Fong and Yildiz, 2007). Under certain growth conditions, however, V. cholerae can also form VPS-independent biofilms (Kierek and Watnick, 2003a,b). This VPS-independent pathway is thought to be preferred in seawater environments, and involves intercellular interactions that occur between Ca+2 and the O-chain of the outer membrane lipopolysaccharide (LPS) (Kierek and Watnick, 2003a).
The regulatory network that controls biofilm formation by regulating expression of vps biosynthesis and matrix protein genes is complex and involves several transcriptional regulators. The core components of this network consist of two positive transcriptional regulators, VpsT and VpsR, and a negative transcriptional regulator HapR (Yildiz and Kolter, 2008). Distruption of vpsT reduces vps gene expression and impedes biofilm formation (Casper-Lindley and Yildiz, 2004). Disruption of vpsR, however, prevents expression of the vps genes and production of VPS, and abolishes formation of typical 3D biofilm structure (Yildiz et al., 2001; Beyhan et al., 2007). Subsequent studies revealed that while VpsR is essential for VPS production and biofilm formation, VpsT plays an accessory role, possibly by increasing the level or activity of VpsR (Beyhan et al., 2007). Finally, HapR, the master quorum sensing regulator, negatively regulates biofilm formation in V. cholerae (Hammer and Bassler, 2003; Zhu and Mekalanos, 2003; Yildiz et al., 2004).
Biofilm formation and dissolution must be tightly controlled to confer a selective advantage to V. cholerae. Some environmental signals should therefore trigger biofilm formation, whereas others should inhibit this process and initiate biofilm dissolution. Natural aquatic ecosystems inhabited by V. cholerae undergo changes in many physicochemical parameters such as nutrient availability, salinity and temperature (Faruque et al., 1998). One of the major environmental fluctuations experienced by V. cholerae is altered Ca2+ levels. The Ca2+ concentration ([Ca2+]) in aquatic environments varies from micromolar levels in freshwater to millimolar (~10 mM) levels in marine environments (Riley and Tongudai, 1967).
The known fluctuations in [Ca2+] in V. cholerae’s aquatic habitats, and the effect of calcium on VPS-independent biofilm formation (Kierek and Watnick, 2003a) led us to investigate effects of Ca2+ on V. cholerae and its VPS-dependent biofilms. In this study, we analysed the response and adaptation of V. cholerae to an external [Ca2+] increase. We determined that Ca2+ negatively regulates transcription of genes that are involved in VPS-dependent biofilm formation. We also identified a two-component regulatory system (now termed calcium-regulated sensor (CarS) and regulator (CarR)), whose transcription decreases in response to an external [Ca2+] increase. Mutational and phenotypic analysis of these regulatory genes revealed that the CarRS two-component regulatory system negatively regulates vps gene expression and biofilm formation in V. cholerae. Epistasis analysis further showed that CarR and HapR act in parallel pathways to negatively regulate biofilm formation in V. cholerae.
To identify genes that are regulated by Ca2+, we compared whole-genome expression profiles of the wild-type V. cholerae cells grown in Luria–Bertani (LB) medium alone and in LB medium supplemented with 10 mM CaCl2 (LBCa2+) to mimic high Ca2+ levels of marine environments. Gene expression data were analysed using the Significance Analysis of Microarrays (SAM) program. We applied the following criteria to define significantly regulated genes: ≤ 3% false discovery rate and ≥ 1.5-fold transcript abundance differences between the samples. Using the selection criteria given above, a total of 76 genes were found to be differentially regulated in cells grown in LBCa2+ compared with the cells grown in LB. Seventeen of these genes were induced and 59 were repressed in Ca2+-adapted cells, as compared with those grown in LB (Table S1). Genes required for different cellular processes including virulence, iron acquisition, biofilm formation and transcriptional regulation were regulated by Ca2+ (Table 1 and Table S1). In this study, however, we focused on two sets of genes: the ones involved in biofilm formation and the ones predicted to be transcriptional regulators.
Ca2+ decreased the message abundance of many of the genes encoding proteins required for VPS biosynthesis (Table 1). VPS biosynthesis genes (vps) are clustered in two regions on the large chromosome of V. cholerae O1 El Tor [vpsU (VC0916), vpsA-K, VC0917-27 (vps-I cluster); vpsL-Q, VC0934-9 (vps-II cluster)]. This result was confirmed by determining β-galactosidase activity in V. cholerae carrying lacZ transcriptional fusions to the promoter of the first genes in the predicted vps-I (vpsA–lacZ) and vps-II operons (vpsL–lacZ). Cells grown in LBCa2+ exhibited a threefold decrease in vpsA–lacZ and fourfold decrease in vpsL–lacZ β-galactosidase activities, relative to cells grown in LB (Fig. 1A). Transcription of the vps genes is positively regulated by the transcriptional regulators VpsT and VpsR (Yildiz et al., 2001; Casper-Lindley and Yildiz, 2004). Analysis of the expression profiling results revealed that the expression of vpsR is decreased in cells grown in LBCa2+ compared with those grown in LB. This result was confirmed by determining β-galactosidase activity in wild-type cells carrying lacZ transcriptional fusions to the promoter of the vpsR; transcription of vpsR decreased by 2.8-fold (Fig. 1A). We determined that transcription of vpsT is also decreased by 2.3-fold in cells grown in LBCa2+ (Fig. 1A). Taken together, these results indicate that decreased vps gene expression in cells grown in LBCa2+ could be due to decreased transcription of the vpsR and vpsT genes.
A major component of V. cholerae biofilm matrix is VPS, which is essential for the development of mature biofilm structures (Watnick and Kolter, 1999; Yildiz and Schoolnik, 1999). Because Ca2+ affected the message abundance of the vps genes, we analysed the effect of Ca2+ on biofilm formation. We compared structure of biofilms, formed by wild-type cells grown in LB to those grown in LBCa2+ using a flow-cell system (Fig. 1B).
Quantitative analysis of these biofilm images revealed that Ca2+ significantly altered biofilm structure (Fig. 1C). While biofilms grown in LB had increased average biofilm thickness, their maximum thickness values were similar (Fig. 1C). Biofilm roughness, a measure of biofilm thickness heterogeneity, was greater for biofilms grown in LBCa2+ compared with biofilms grown in LB, indicating that the surface architecture of the LB biofilms is more regular (Fig. 1C).
We reasoned that the stability of biofilms grown in LB and LBCa2+ could also be different. To test this possibility, we treated 48-hour-old biofilms grown in LBCa2+ and in LB with 0.5% SDS for 15 min, and determined biofilm structures. SDS addition led to higher dissolution in biofilms grown in LBCa2+, indicating that biofilms grown in LBCa2+ were less stable compared with biofilms grown in LB (Fig. 1B). Lower vps gene expression at high [Ca2+] –resulting in a decrease in VPS production – is the likely cause of this instability. An alternative explanation is that interactions between Ca+2 and the O-chain of the outer membrane LPS or biofilm matrix proteins lead to changes in biofilm stability.
To evaluate the importance of VPS to Ca2+-mediated changes to biofilm structure and stability, we tested the effect of an increased [Ca2+] on biofilms of a mutant that is unable to produce VPS, due to deletion of vps-I and vps-II clusters (Δvps-IΔvps-II, called Δvps). Biofilms of the Δvps mutant consisted of flat monolayers of cells; thus, the effect of Ca2+ on biofilm thickness and roughness due to VPS production cannot be evaluated. Hence, we looked at SDS mediated detachment of biofilms. Biofilms of Δvps mutants detached readily from the substratum upon treatment with 0.5% SDS (Fig. 1D). These results indicate that VPS is necessary for biofilm stability. Interestingly, however, Ca2+ addition to Δvps biofilms led to detachment, decrease in substratum coverage and formation of clusters of microcolonies (Fig. 1D and E). Such aggregates were also present in wild-type biofilms grown in LBCa2+ (Fig. 1B). Together, these observations suggest that Ca2+ addition alters either cell-cell or cell–surface interactions to create more clustered but less well attached biofilms. This finding also suggests that there are other factors besides VPS that are involved in calcium-mediated changes in biofilm structure. This observation is consistent with the report that V. cholerae can form VPS-independent biofilms that require interactions between Ca+2 and the O-chain of the outer membrane LPS (Kierek and Watnick, 2003a).
We also analysed the effect of [Ca2+] on preformed biofilms in two different ways. In the first set of flow-cell experiments, biofilms were grown in LB for 24 h then the media was supplemented with 10 mM Ca2+, and biofilms were analysed after 24 h of incubation (Fig. S1A). Quantitative analysis of these biofilm images revealed that Ca2+ significantly altered the architecture of biofilms (Fig. S2A). Ca2+ addition to preformed biofilms decreased the average biofilm thickness and increased biofilm roughness. It is noteworthy that the analysis of the effect of Ca2+ addition on biofilms grown in defined artificial sea water-based medium also revealed similar results (Fig. S1C). In the second set of experiments, we analysed the effect of Ca2+ removal on biofilms grown in LBCa2+ for 24 h (Fig. S1B). In this case, while the mean thicknesses decreased, surface coverage dramatically increased in the biofilms 24 h after the removal of Ca2+ from the biofilm medium (Fig. S2B). This decrease in biofilm thickness is most likely due to the increased compactness of biofilms with an increase in VPS and biofilm matrix protein productions. Collectively, these studies clearly showed that the structural properties of biofilms formed in LBCa2+ are different than those formed in LB.
Transcriptional regulators, VpsR and VpsT positively regulate VPS production and biofilm formation. Because our microarray data above suggested that [Ca2+] affects vps gene expression, we wanted to further analyse how VpsR and VpsT are involved in Ca2+-mediated repression of vps gene expression. To this end, we introduced a plasmid harbouring a vpsL–lacZ transcriptional fusion into ΔvpsR and ΔvpsT mutants, and measured vpsL transcription in cells grown in LB and LBCa2+. No vpsL transcription was detected in vpsR mutant cells grown in either LB or LBCa2+, as expected (Fig. 2). The vpsT mutant exhibited a 3.7-fold decrease in vpsL expression, compared with that of the wild-type cells grown in LB (Fig. 2). Wild-type cells showed an approximately fivefold difference of vpsL expression between cells grown in LB or LBCa2+, the vpsT mutant showed a threefold difference, and the vpsR mutant showed essentially no vpsL expression. This finding suggests that the ΔvpsT strain is still able to respond to an increase in external Ca2+ levels.
The quorum sensing regulator, HapR, was previously shown to negatively regulate transcription of vps genes (Hammer and Bassler, 2003; Zhu and Mekalanos, 2003; Yildiz et al., 2004). To determine whether this regulator was important for Ca2+-mediated repression of vps gene expression, we also analysed vpsL expression in a ΔhapR mutant grown in LB and LBCa2+. Similar to the wild-type, vpsL expression decreased in cells grown in LBCa2+ (Fig. 2), indicating that HapR does not regulate vps expression in a Ca2+-dependent manner.
Comparison of whole-genome expression profiles of the cells grown in LBCa2+ to cells grown in LB medium revealed that Ca2+ leads to a decrease in transcription of genes predicted to encode proteins with regulatory functions (i.e. VC1319, encoding a sensor histidine kinase, and VC1320, encoding a response regulator). Message abundance of VC1319 and VC1320 were decreased by 1.6-fold and 2.3-fold respectively (Table 1 and Table S1). We also made the same observation when we analysed the initial response of V. cholerae to an increase in external [Ca2+] by monitoring changes in the gene expression profile in exponentially grown cells that were subjected to ‘Ca2+ addition’ for 15 and 30 min (Table S1). We named VC1319 ‘carS’ for calcium regulated sensor histidine kinase and VC1320 ‘carR’ for calcium regulated response regulator.
The carS gene is predicted to encode a 440-aminoacid, 50.2 kDa protein that exhibits 27.1% identity/46.0% similarity to Escherichia coli RstB. The carR gene is predicted to encode a 234-amino-acid, 26.4 kDa protein that exhibits 44.6% identity/63.6% similarity to E. coli RstA. The RstAB two-component regulatory pair is part of the divalent cation-sensing PhoPQ regulon (Minagawa et al., 2003; Ogasawara et al., 2007); however, the environmental stimuli required for the activation of RstAB are unknown (Eguchi et al., 2004).
To confirm the expression profile data, we constructed carR–lacZ transcriptional fusion (carS and carR are separated by 25 bp and predicted to be organized in an operon where carR is the first gene) by amplifying the upstream regulatory sequence of carR, and inserting it into the upstream of a promoterless lacZ gene in vector pRS415. Transcription was measured by determining β-galactosidase activity in wild-type cells grown in LB and LBCa2+. Cells grown in LBCa2+ exhibited a decrease in β-galactosidase activity relative to cells grown in LB (Fig. 3A), confirming the trend of the microarray experiment.
As we determined that expression of both vps and carRS genes are negatively regulated in cells treated with Ca+2, we wanted to determine whether CarRS is involved in Ca2+-mediated repression of vps genes. To this end, we generated in-frame deletion mutants of carS and carR. To identify genes regulated by CarS and CarR, we first compared whole-genome expression profiles of the ΔcarS and ΔcarR mutants to the wild-type using total RNA isolated from cells grown in LB. Gene expression data were analysed using the SAM program with the criteria discussed above. We found a total of 19 differentially regulated genes in ΔcarS compared with the wild-type during exponential phase. Of these genes, 9 were induced and 10 were repressed in ΔcarS compared with the wild-type (Table S1 and Fig. S3). In the ΔcarR mutant, 67 genes were differentially expressed; of which 27 were induced and 40 were repressed compared with the wild-type (Table S1 and Fig. S3).
Interestingly, expression profiling showed that mRNA abundance of many of the vps and vps intergenic region genes were increased in the ΔcarR mutant by 1.5- to 2.7-fold relative to the wild-type. Similarly, mRNA levels from some of the vps genes (VC0918, VC0933 and VC0935) were also increased in the ΔcarS mutant relative to the wild-type. This result was confirmed by determining β-galactosidase activity in strains harbouring plasmids with vpsA–lacZ and vpsL–lacZ transcriptional fusions. Transcription of both vpsA and vpsL was increased in the ΔcarS and ΔcarR mutants relative to the wild-type (Fig. 3B). Taken together, these results suggest that CarR and CarS negatively regulate vps gene expression.
To better evaluate the mechanism by which CarR negatively regulates vps expression, we analysed vpsT and vpsR transcription in the wild-type and ΔcarR strains harbouring vpsT–lacZ and vpsR–lacZ fusion plasmids (Fig. 3C). A 2- and 1.4-fold increase in β-galactosidase activities was observed in ΔcarR harbouring vpsT–lacZ and vpsR–lacZ fusion plasmids, respectively, when compared with the wild-type. These results suggest that CarR negatively regulates the expression of both vpsT and vpsR, and that increased vps expression in ΔcarR could be due to increased transcription of the vpsR and vpsT genes.
To further evaluate contributions of CarS and CarR to Ca2+ adaptation response, we also determined the whole-genome transcriptional profile of ΔcarS and ΔcarR mutants upon an increase in [Ca2+]. In large part, Ca2+ adaptation responses of these mutants were similar to those of the wild-type (Fig. S3). This observation indicates that the CarRS regulatory system regulates expression of only a small set of genes that is regulated by external [Ca2+] increase.
To confirm that the CarRS system is not involved in the Ca2+ regulation of the vps genes, we compared vpsA and vpsL expressions in the wild-type, ΔcarS and ΔcarR strains grown in LB and LBCa2+. Expression of vpsA and vpsL genes was higher in ΔcarS and ΔcarR mutants relative to wild-type when grown in LB or LBCa2+. However, we found decreased β-galactosidase activities in the wild-type, ΔcarS and ΔcarR strains grown in LBCa2+, relative to cells grown in LB (Fig. 3D). These results are consistent with microarray results, showing that the CarRS two-component regulatory system is not involved in Ca2+-mediated repression of vps gene expression.
Taken together, our results suggest that in Ca2+-treated cells there appears to be two parallel signalling pathways controlling vps gene expression. In the first pathway transcription of vpsR and vpsT, and in turn vps structural genes, is decreased by a yet to be determined mechanism. The second pathway is controlled by CarR, where a decrease in the transcription of carR would lead to an increase in the transcription of vpsR, vpsT and vps genes (Fig. 3). Because Ca2+-treated wild-type cells have decreased vps expression, the first pathway appears to be dominant over CarRS-dependent pathway.
CarS and CarR negatively regulate transcription of vps genes; therefore, ΔcarS and ΔcarR mutants are expected to have altered biofilm structures. To test this idea, we compared the biofilms of the ΔcarS and ΔcarR mutants to the wild-type biofilms. We first analysed biofilms formed on chambered coverglasses after 8 h of incubation under static conditions at 30°C, with biofilm structures analysed by confocal scanning laser microscopy (CSLM). The results, shown in Fig. 4A, revealed that both ΔcarS and ΔcarR mutants have an enhanced capacity to form biofilms. To gain further insight into the structure of biofilms formed by the ΔcarS and ΔcarR mutants, biofilms were grown at room temperature in flow cells for 48 h (Fig. 4B). Biofilms formed by the ΔcarS and ΔcarR mutants had higher average thickness and had more pillar-like structures relative to the wild-type biofilms (Fig. 4B and Table 2).
To further evaluate the effect of CarS and CarR on biofilm formation, we overexpressed carS and carR from an arabinose-inducible vector in wild-type cells (Fig. 4C). Biofilms of cells harbouring carR and carS overexpression plasmids were grown in a flow-cell system, using LB medium supplemented with 0.2% arabinose, with images acquired by CSLM over the course of the biofilm development (Fig. 4C). There was a significant difference in biofilm formation dynamics in cells overexpressing CarR, relative to the control strain. In cells overexpressing CarR, surface attachment and surface coverage were decreased dramatically 8 and 24 h after the inoculation (Fig. 4C and Table 2). Although strains overexpressing CarS exhibited a decrease in surface coverage, the defect was less profound than that seen with CarR overexpression. After 48 h of biofilm development, strains overexpressing CarS and CarR were able to form biofilms. However, final biofilm architecture and surface colonization were different than in the biofilms of the control strain. Taken together, our studies indicate that the CarRS regulatory system negatively regulates biofilm formation in V. cholerae.
VpsR is the most downstream regulator of vps gene transcription in the VpsT, VpsR and HapR regulatory circuitry (Beyhan et al., 2007). To determine how CarR contributes to this regulatory circuitry, we generated ΔcarRΔvpsR, ΔcarRΔvpsT and ΔcarRΔhapR double mutants. We then monitored transcription of the vpsL–lacZ fusion in wild-type, ΔcarR, ΔvpsT, ΔvpsR, ΔcarRΔvpsT and ΔcarR-ΔvpsR mutants (Fig. 5A). As discussed above, in ΔcarR mutant the transcription of vpsL is markedly increased compared with the wild-type. Transcription of vpsL–lacZ is decreased in ΔvpsT and ΔvpsR by 2.5-fold and 57.5-fold, respectively, compared with the wild-type (Fig. 5A). While vpsL expression was similar in the ΔvpsR and ΔcarR-ΔvpsR strains, indicating that VpsR acts downstream of CarR. vpsL expression was 2.6-fold higher in the ΔcarR-ΔvpsT mutant relative to that of the ΔvpsT mutant, indicating that VpsT and CarR act on parallel pathways. We reported previously that VpsT positively regulates transcription of vpsR (Casper-Lindley and Yildiz, 2004). Hence, the decrease in vpsL transcription in the ΔcarR-ΔvpsT strain is likely due to decreased expression of vpsR.
Biofilm formation in V. cholerae is negatively regulated by HapR. To determine a possible connection between CarR and HapR, we monitored transcription of vpsL, using lacZ transcriptional fusion constructs in wild-type, ΔcarR, ΔhapR and ΔcarRΔhapR mutants (Fig. 5A). Transcription of vpsL–lacZ was increased in ΔcarR and ΔhapR by 3- and 10-fold, respectively, compared with the wild-type (Fig. 5A). Transcription of vpsL–lacZ was higher in the ΔcarRΔhapR strain than in the ΔhapR strain, indicating that CarR and HapR act in parallel pathways to negatively regulate vps gene expression (Fig. 5A).
In bacteria, relatively little is known about Ca2+ signalling and the processes that are regulated by Ca2+ (reviewed in Norris et al., 1991; Norris et al., 1996; Dominguez, 2004). Intracellular [Ca2+] in bacterial cells is dynamic, and changes in intracellular [Ca2+] could regulate diverse cellular processes (reviewed in Norris et al., 1991; Norris et al., 1996; Dominguez, 2004). Intracellular [Ca2+] measurements in E. coli revealed that [Ca2+] ranges from 170 to 300 nM (Jones et al., 1999). Furthermore, intracellular Ca2+ levels were found to vary as a function of external Ca2+ levels, reaching approximately 2 μM upon exposure to 1 mM external [Ca2+] (Holland et al., 1999). Intracellular Ca2+ levels in bacteria are controlled by influx mechanisms through the actions of ion channels, primary and secondary transporters, Ca2+ export systems and Ca2+-binding proteins (CaBP) (Norris et al., 1996; Jones et al., 1999). Because V. cholerae is predicted to experience [Ca2+] fluxes in its environment, we reasoned that it might respond to changes in external [Ca2+]. In this study, we determined that an increase in [Ca2+] leads to a decrease in vps gene expression and VPS-dependent biofilm formation in V. cholerae.
Ca2+ has been shown to influence surface attachment and biofilm formation in other bacteria. The role of Ca2+ can be due to its bridging role in the extracellular polymeric matrix (Rose and Turner, 1998; Rose, 2000), its interactions with extracellular or cell surface-associated Ca2+-binding proteins (Arrizubieta et al., 2004), or its regulatory effects on the expression of genes known to play roles in biofilm formation and surface attachment. A proteomic analysis, designed to elucidate cellular response to an increase in external [Ca2+] in the marine bacterium Pseudoalteromonas sp. 1398, revealed that an increase in [Ca2+] caused a global change in the protein expression profile during both planktonic and biofilm-mode of growth (Patrauchan et al., 2005). Furthermore, an increase in [Ca2+] was found to enhance extracellular matrix production. In Pseudomonas aeruginosa, an increase in [Ca2+] induces transcription of alginate biosynthesis genes and the amount of alginate in biofilm matrix (Sarkisova et al., 2005). In contrast, we observed that in V. cholerae VPS-dependent extracellular matrix production is decreased in response to an increase in [Ca2+]. In Staphylococcus aureus strains, harbouring the biofilm-associated protein Bap, elevated extracellular calcium levels inhibited intercellular adhesion and biofilm formation (Arrizubieta et al., 2004). Bap is a staphylococcal surface protein that is involved in biofilm formation and binds Ca2+ via an EF-hand motif (Gotz, 2002). It was shown that Bap binds Ca2+ with low affinity and, upon Ca2+ binding, Bap loses its ability to promote biofilm formation (Arrizubieta et al., 2004). Our lab has recently identified a set of biofilm matrix proteins, and two of these proteins, RbmC and Bap1, also harbour a Ca2+-binding domain (Fong and Yildiz, 2007). We are currently testing whether these proteins bind Ca2+ and how these interactions affect V. cholerae biofilm formation.
The mechanism by which V. cholerae senses external [Ca2+] is not known. In Salmonella, a two-component regulatory system, PhoPQ, is required for sensing divalent cations (Ca2+, Mg2+, Mn2), antimicrobial peptides and acidity. The PhoPQ system is active at low divalent cation concentrations, and under such conditions the membrane-localized sensor histidine kinase, PhoQ, auto-phosphorylates and then phosphorylates its cognate response regulator, PhoP (reviewed in Groisman, 2001; Prost and Miller, 2008). The phosphorylated PhoP activates transcription of several genes required for virulence. Vibrio cholerae genome analysis revealed that there are putative PhoPQ homologues (VCA1104-05 and VC1638-39) in the genome, which may be involved in sensing divalent cations, and we are currently analysing their roles in Ca2+ sensing.
In this study, we also identified a set of transcriptional regulators, CarS and CarR, whose expression is down-regulated under increased [Ca2+] conditions. We determined that CarR and CarS negatively regulates vps gene expression and biofilm formation in V. cholerae. These proteins are homologous to RstB and RstA of E. coli respectively. In E. coli, the rstAB operon was originally identified as part of the PhoPQ two-component regulatory system. Because CarRS is homologous to RstAB, our findings together suggest that a PhoPQ-like signal transduction system is also operational in V. cholerae. Phenotypic analysis of the rstAB mutant, using phenotype microarrays, revealed that RstAB is involved in resistance to antibiotics ketoprofen, pridinol and troleandomycin (Zhou et al., 2003). To date, two genes, asr and csgD, have been shown to be under the control of RstA (Ogasawara et al., 2007). asr is needed for the survival of E. coli in a low-pH environment (Seputiene et al., 2003). The second gene that is regulated by RstA is csgD, encoding a transcription factor that is required for the production of biofilm matrix components in E. coli and Salmonella spp. (Romling et al., 1998). RstAB negatively regulates CsgD expression and therefore biofilm formation in E. coli (Ogasawara et al., 2007). Our results showed that in V. cholerae the CarRS two-component system negatively regulates the expression of vpsT, which exhibits 44% sequence identity to CsgD, the positive regulator of vps gene expression and biofilm formation (Casper-Lindley and Yildiz, 2004). Our analysis also revealed that CarRS positively regulates transcription of genes predicted to be involved in lipid A modification (Table S1). It is not known if RstA is involved in such a response, but PhoPQ regulates structural modifications to lipid A in Salmonella spp., and such modifications play a role during infection, mediating resistance to host antimicrobial peptides and avoidance of immune system.
In this study, we showed that extracellular [Ca2+] is an environmental signal that negatively affects VPS-dependent biofilm formation and identified a calcium-controlled negative regulatory system affecting V. cholerae biofilm formation. Better understanding of responses of V. cholerae to the physical and chemical factors that are likely to be encountered by the pathogen in natural aquatic habitats, as well as the environmental signals and regulatory networks that govern transitions between planktonic and biofilm states of V. cholerae in aquatic ecosystems, will shed light on the mechanism of survival of the organism in the environment and transmission to human host.
Bacterial strains and plasmids used in this study are listed in Table 3. Luria–Bertani broth (1% Tryptone, 0.5% Yeast Extract and 1% NaCl) at pH 7.0 or LB supplemented with 10 mM CaCl2 (LBCa2+) at pH 7.0 were used. Vibrio cholerae cultures were grown in LB or LBCa2+ at 30°C to an OD600 of 0.3 (corresponding to ~2 × 108 cells ml−1) in 125 ml flasks containing 25 ml of growth medium and were shaken at 200 r.p.m. Escherichia coli cultures were grown at 37°C in LB and were shaken at 200 r.p.m. Under this condition, growth kinetics of V. cholerae cells grown in LB or LBCa2+ are nearly identical (data not shown). Unless otherwise noted, the antibiotics rifampicin (Rif) and ampicillin (Amp) were added at concentrations of 100 μg ml−1, and gentamicin was added at concentration of 50 μg ml−1. LB broth without NaCl and with 10% sucrose was used for counter selection with sacB-containing plasmids. Arabinose, 0.2% (w/v), was used for induction of gene expression and was added to the growth medium when necessary. The salt base of ASW used in biofilm assays contained 9.25 mM CaCl2, 8.32 mM KCl, 23.12 mM MgCl2·6H2O, 52.34 mM MgSO4, 422.66 mM NaCl, 2.14 mM NaHCO3, 0.07 mM NaF, 0.747 mM KBr, 0.388 mM H3BO3 and 0.15 mM SrCl2. The (ASW) salt base was supplemented with 0.005% K2HPO4, 0.1% NH4Cl, 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (Hepes), 1× MEM Vitamin Solution (Gibco-Invitrogen), 0.2% peptone and 0.01% yeast extract.
All oligonucleotides used for PCR analysis and DNA sequencing were obtained from Operon Technologies (Alameda, CA) and are listed in Table S2. All PCRs were performed with the High-Fidelity or Expand High-Fidelity PCR kits (Roche). PCR products and plasmids were cleaned and prepared using QiaQuick PCR purification (QIAGEN), QiaPrep Spin Miniprep (QIAGEN) and GFX PCR DNA Gel Band Purification (GE Healthcare) kits. DNA sequencing was done at the UC Berkeley DNA Sequencing Facility.
Deletions for all genes were carried out using the same general strategy, as described in Lim and colleagues (2006). Briefly, a ~600 bp fragment 5′ of the gene, including several nucleotides of the gene, was amplified by PCR with primers A and B (Table S2). A similar fragment was also amplified from the 3′ end of the gene, using primers C and D (Table S2). Purified PCR fragments from these reactions were allowed to anneal to sequences in primers B and C and amplified in a second PCR reaction. The resulting ~1200 bp fragment was then amplified with primers A and D, creating the deletion construct. The purified PCR fragment was digested with two of the following restriction enzymes: SacI, NcoI or XbaI, and ligated into a pGP704-sacB28 suicide plasmid digested with the same enzymes. The deletion constructs are listed in Table 3. The deletion plasmids were maintained in E. coli CC118 (λ pir). Biparental matings were carried out with the wild-type V. cholerae and a conjugative strain E. coli S17-1 (λ pir) harbouring deletion plasmid. Selection of deletion mutants were done as described by Fong and Yildiz (2007) and were verified by PCR. The gfp-tagged V. cholerae strains were generated by triparental matings with donor E. coli S17-1 (λ pir) carrying pMCM11, helper E. coli S17-1 (λ pir) harbouring pUX-BF13 and V. cholerae strains. Transconjugants were selected on thiosulfate-citrate-bile salts-sucrose (Difco) agar medium containing gentamicin at 30°C. The gfp-tagged V. cholerae strains were verified by PCR.
Total RNA was isolated from V. cholerae cells grown in LB medium to an OD600 of 0.3–0.4. RNA isolation was done as described by Beyhan and colleagues (2007). To analyse the initial response to an increase in external [Ca+2], a final concentration of 10 mM Ca2+ was added to exponentially grown cells and cells were collected 15 and 30 min after Ca2+ addition. To analyse adaptation to an increase in external [Ca+2], overnight grown cultures of V. cholerae in LB medium at 30°C were diluted 1:200 in LBCa2+ and incubated at 30°C with shaking (200 r.p.m), until they reached to an OD600 of 0.3–0.4. To ensure homogeneity, these cultures were diluted again 1:200 in LBCa2+ medium and grown to an OD600 of 0.3–0.4. Aliquots of 1.8 ml were collected by centrifugation for 2 min at room temperature. The cell pellets were immediately resuspended in 1 ml of Trizol reagent (Invitrogen) and stored at −80°C. The total RNA from the pellets was isolated according to the manufacturer’s instructions. To remove contaminating DNA, total RNA was incubated with RNase-free DNase I (Ambion), and the RNeasy Mini kit (QIAGEN) was used to clean up the RNA after DNase digestion.
Microarrays used in this study were composed of 70-mer oligonucleotides, representing the open reading frames present in V. cholerae genome. These were designed and synthesized by Illumina, San Diego, and printed at UCSC. Whole-genome expression analysis was performed using a common reference, which was RNA-isolated from wild-type strains grown to an OD600 of 0.3–0.4. RNA samples from test and reference samples were used in cDNA synthesis, and microarray hybridization and scanning were performed as described previously (Beyhan et al., 2006; 2007). Normalized signal ratios were obtained with LOWESS print-tip normalization, using the Bioconductor packages (http://www.bioconductor.org) in an R environment. Differentially regulated genes were determined (with 3 biological and 2 technical replicates for the analysis of Ca2+ adaptation and ΔcarR/S regulation or, 2 biological and 2 technical replicates for identification of initial Ca2+ response and the ΔcarR/S response to Ca2+) using the SAM software (Tusher et al., 2001), with a 1.5-fold difference in gene expression and a 3% false discovery rate as a cutoff value. Identification of genes with similar expression pattern was performed by analysing the data using GENESIS software (Sturn et al., 2002).
Overnight grown cultures of V. cholerae in LB medium at 30°C were diluted 1:200 in either LB or LBCa2+ and incubated at 30°C with shaking (200 r.p.m), until they reached to an OD600 of 0.3–0.4. To ensure homogeneity, these cultures were diluted again 1:200 in appropriate medium and grown to an OD600 of 0.3–0.4. Cell harvesting and β-galactosidase assays were carried out according to a previously published procedure (Fong et al., 2006). The assays were repeated with three different biological replicates and at least eight technical replicates.
Biofilms were grown either at room temperature in flow cells (individual channel dimensions of 1 × 4 × 40 mm) or at 30°C in chambered cover-slides. Biofilms in flow cells were supplied with 2% LB (0.01% yeast extract, 0.02% tryptone and 1% NaCl) or ASW at a flow rate of 10.2 ml h−1. The flow-cell system was assembled and prepared as described previously (Heydorn et al., 2000). Cultures for inoculation of the flow cells were prepared by inoculating a single colony from a plate into flasks containing LB medium and growing them with aeration at 30°C for 16 h. Cultures were diluted to an OD600 of 0.1 in 2% LB and used for inoculation. A 300 μl volume of diluted culture was injected into each chamber with a small syringe. After inoculation, flow cells were left at room temperature for 1 h without flow. The flow was then started at a constant rate of 10.2 ml h−1 with a Watson Marlow 205S peristaltic pump. Static biofilms were formed in chambered cover-slides at 30°C. Cultures for inoculation of the chambered cover-slides were grown as described above. Cultures were diluted to an OD600 of 0.02 in LB and 3 ml of this dilution was used for inoculation. Three millilitres of 0.5% SDS in 0.9% NaCl was administered into flow cells at 48 h time point to test biofilm stability. Biofilms were incubated for 15 min without flow, and then the flow reinstated and biofilms were incubated for another 15 min in the presence of appropriate growth medium. Confocal scanning laser microscopy images of the biofilms were captured with a LSM 5 PASCAL system (Zeiss) at 488 nm excitation and 543 nm emission wavelengths. Three-dimensional images of the biofilms were reconstructed using Imaris software (Bitplane). Images from each experiment were analysed using the computer program COMSTAT (Heydorn et al., 2000) in MATLAB environment.
Fig. S1. Effect of Ca2+ on biofilm formation.
A. Changes in the architecture of the biofilms in response to Ca2+ addition in the gfp-tagged wild-type cells. Biofilms of the wild-type cells were grown in flow cells for 24 h in LB (24 h –LB); 10 mM Ca2+ was then added to the growth medium and biofilms were allowed to develop for an additional 24 h (48 h – Ca2+ addition). Control biofilms were grown in flow cells without Ca2+ for 48 h (48 h – LB). 0.5% SDS was administered into each chamber at a 48 h time point to test biofilm stability. Images were taken in two magnifications: 40× where white bars represent 30 μm and 20× (marked with an asterisk) where white bars represent 50 μm. Images were acquired using a CSLM and processed using Imaris software. The result shown is representative of three independent experiments.
B. Changes in the architecture of the biofilms in response to Ca2+ removal in the gfp-tagged wild-type cells. Biofilms were grown in flow cells for 24 h in LBCa2+ (24 h – LBCa2+), then growth medium was changed to LB and biofilms were allowed to develop for an additional 24 h (48 h – Ca2+ removal). Control biofilms were grown in flow cells for 48 h in LBCa2+ (48 h – LBCa2+). 0.5% SDS was administered into each chamber at the end of 48 h incubation period to test biofilm stability. Images were taken in two magnifications: 40× where white bars represent 30 μm and 20× (marked with an asterisk) where white bars represent 50 μm. Images were acquired using a CSLM and processed using Imaris software. The result shown is representative of three independent experiments.
C. Changes in the architecture of the biofilms in response to Ca2+ addition in gfp-tagged wild-type cells. Biofilms of wild-type cells were grown in flow cells for 24 h in artificial sea water (ASW) supplemented with 0.5 mM Ca2+ (24 h – ASW [0.5 mM Ca2+]); the growth medium was then switched to ASW supplemented with 10 mM Ca2+ and biofilms were allowed to develop for an additional 24 h (48 h – Ca2+ addition). Control biofilms were grown in flow cells for 48 h in ASW supplemented with 0.5 mM Ca2+ (48 h – ASW [0.5mM Ca2+]). 0.5% SDS was administered into each chamber at a 48 h time point to test biofilm stability. Images were acquired using a CSLM and processed using Imaris software. The result shown is the representative of three independent experiments. White bars represent 30 μm.
Fig. S2. Quantitative analysis of biofilms. Changes in the structural characteristics of biofilms that were formed by the wild-type cells in response to Ca2+ addition (A) and removal (B) were quantified using COMSTAT at 48 hour time point. Results shown were calculated from three image stacks that were representative of three independent experiments.
Fig. S3. Gene expression profiles of ΔcarS and ΔcarR mutants grown in LBCa2+ and LB.
Table S1. Genes that are differentially expressed in wild-type grown in LBCa2+, and wild-type, ΔcarS and ΔcarR grown in LB.
Table S2. Primers used in this study.
This work was supported by grants from NIH (RO1 AI055987) and CEQI0047 provided by the UC Marine Council Coastal Environmental Quality Initiative. We thank Karen Ottemann, Chad Saltikov, Manel Camps and members of the Yildiz laboratory for their suggestions.